I was bored and found a HP-27S

[jebem]

As I am in the tropics now working this weeks on a project for one of our VMware partners, and with nothing to do but browsing the net at the evenings, I found a HP-27S for sale at an irresistible price, despite showing some wear in the front display area.
So I bought it right away.

As this model is overlooked in favor of the traditional RPN models, I decided that it was time to have this sort of algebraic HP calculator in my collection for a change.

Apparently its set of do it all features are enough to take on most of the other HP calculators of its time, almost as powerful as the 42S on scientific features, and showing the shortcomings of machines like the 32S series.

Its only sin is that it can't RPN... despite this, I see the high asking prices, perhaps because it is a rare machine these days?

[Don Shepherd]

(06-09-2014 10:38 PM)jebem Wrote:  showing the shortcomings of machines like the 32S series.

Shortcomings of 32s? I'll take programmability over TVM any day.

[Katie Wasserman]

(06-09-2014 11:37 PM)Don Shepherd Wrote:  Shortcomings of 32s? I'll take programmability over TVM any day.

You know I love the 32s/ii and programming it, but I have to say that I bought a 27s when it first came out and I loved playing with it, it's not just TVM

It has a the solver with L() and G() that work -- same as the original 17b/ii, access to the various list-based items, trig functions, appointments, base conversion, printing. It's quite nice, except for lacking RPN.

[rprosperi]

(06-10-2014 02:10 AM)Katie Wasserman Wrote:  

(06-09-2014 11:37 PM)Don Shepherd Wrote:  Shortcomings of 32s? I'll take programmability over TVM any day.

You know I love the 32s/ii and programming it, but I have to say that I bought a 27s when it first came out and I loved playing with it, it's not just TVM

It has a the solver with L() and G() that work -- same as the original 17b/ii, access to the various list-based items, trig functions, appointments, base conversion, printing. It's quite nice, except for lacking RPN.

I agree with Katie, a very nice machine. I would say its the nicest algebraic machine HP has made (excluding RPL machines which some folks classify as Algebraic, despite the inherent sin of doing so). It has all the key features of the 19B/19BII, but in a Pioneer form factor (Who ever used that Text notepad feature in the 19B anyhow?)

They seem to have been somewhat rare, but they do show up on TAS often, though frequently in poor shape.

[John R]

(06-09-2014 10:38 PM)jebem Wrote:  Apparently its set of do it all features are enough to take on most of the other HP calculators of its time, almost as powerful as the 42S on scientific features, and showing the shortcomings of machines like the 32S series.

The 27S definitely seems like a hybrid of the 17B and 42S, with maybe 60% of its DNA from the former and 40% from the latter. Which means its numerical designation is particularly fitting, since (0.6 * 17) + (0.4 * 42) = 27.

[walter b]

(06-10-2014 04:39 AM)John R Wrote:  The 27S definitely seems like a hybrid of the 17B and 42S, with maybe 60% of its DNA from the former and 40% from the latter. Which means its numerical designation is particularly fitting, since (0.6 * 17) + (0.4 * 42) = 27.

AFAIK the 27S came before the 17B. But it was a nice theory ...

[Didier Lachieze]

(06-10-2014 05:55 AM)walter b Wrote:  

(06-10-2014 04:39 AM)John R Wrote:  The 27S definitely seems like a hybrid of the 17B and 42S, with maybe 60% of its DNA from the former and 40% from the latter. Which means its numerical designation is particularly fitting, since (0.6 * 17) + (0.4 * 42) = 27.

AFAIK the 27S came before the 17B. But it was a nice theory ...

Well, they were both introduced in January 1988 at the Winter Consumer Electronic Show, along with the 19BII and the 28S. (but the 42S was introduced after the 27S, on October 31, 1988).

Extract from the HP Press Release:

Quote:FOUR NEW HP CALCULATORS EXPAND PRODUCT OFFERING

Hewlett-Packard Company is using the newest automated-production techniques to carve out a successful position for itself in the consumer-electronic market, long dominated by the Japanese.

It's latest offerings in the highly competitive calculator industry--four new handheld models--will be introduced in Las Vegas beginning on Jan. 7.

The new products include:
o HP-19B Business Consultant II;
o HP-17B business calculator;
o HP-28S advanced scientific calculator; and
o HP-27S scientific calculator.
[...]

[walter b]

(06-10-2014 06:37 AM)Didier Lachieze Wrote:  

(06-10-2014 05:55 AM)walter b Wrote:  AFAIK the 27S came before the 17B. But it was a nice theory ...

Well, they were both introduced in January 1988 at the Winter Consumer Electronic Show, along with the 19BII and the 28S. (but the 42S was introduced after the 27S, on October 31, 1988).

Merci. I stand corrected.

[jebem]

(06-09-2014 11:37 PM)Don Shepherd Wrote:  

(06-09-2014 10:38 PM)jebem Wrote:  showing the shortcomings of machines like the 32S series.

Shortcomings of 32s? I'll take programmability over TVM any day.

OK, good point!
The HP-27S lacks conventional programmability and that is a shortcoming for me as well.
But for the price I have paid for it - just 28 Euros + s&h, and considering that it is in good working condition (well, at least it was advertized as such) and despite showing some scratches in the display frame, I still believe it was a nice acquisition

When I receive it at home and when I'm back to Lisbon, I will report again.

[Don Shepherd]

(06-10-2014 09:35 AM)jebem Wrote:  

(06-09-2014 11:37 PM)Don Shepherd Wrote:  Shortcomings of 32s? I'll take programmability over TVM any day.

OK, good point!
The HP-27S lacks conventional programmability and that is a shortcoming for me as well.
But for the price I have paid for it - just 28 Euros + s&h, and considering that it is in good working condition (well, at least it was advertized as such) and despite showing some scratches in the display frame, I still believe it was a nice acquisition

When I receive it at home and when I'm back to Lisbon, I will report again.

Sounds like a good deal. Katie reminded me about the 27s solver programming, I had forgotten about that. In my opinion, one of HP's greatest achievements was the inclusion of programming constructs such as IF, sigma (loop), and L() and G() in the solver found in several calculator models, including the 27s. This made the solver more than just an equation solver; it is a real minimum programming language.

HP should have received some type of award for this solver, but we are all beneficiaries.

[Katie Wasserman]

(06-10-2014 11:18 AM)Don Shepherd Wrote:  In my opinion, one of HP's greatest achievements was the inclusion of programming constructs such as IF, sigma (loop), and L() and G() in the solver found in several calculator models, including the 27s. This made the solver more than just an equation solver; it is a real minimum programming language.

HP should have received some type of award for this solver, but we are all beneficiaries.

If I recall correctly the HP "step-by-step" publication called HP-27S/19B Technical Applications introduced L(), G() and other solver techniques. This was available when the 27S and 19B were first released. It showed how you could use the solver to compute coordinated transforms, solve linear equations, do complex variable arithmetic, find GCD and factor prime among other things. Pretty amazing for a "non-programmable" calculator.

[Thomas Radtke]

(06-10-2014 03:52 PM)Katie Wasserman Wrote:  Pretty amazing for a "non-programmable" calculator.

It is programmable, the solver is just not turing complete. Calculators like the 20S play in the same league, allthough not as obvious.

[rprosperi]

(06-10-2014 03:52 PM)Katie Wasserman Wrote:  

(06-10-2014 11:18 AM)Don Shepherd Wrote:  In my opinion, one of HP's greatest achievements was the inclusion of programming constructs such as IF, sigma (loop), and L() and G() in the solver found in several calculator models, including the 27s. This made the solver more than just an equation solver; it is a real minimum programming language.

HP should have received some type of award for this solver, but we are all beneficiaries.

If I recall correctly the HP "step-by-step" publication called HP-27S/19B Technical Applications introduced L(), G() and other solver techniques. This was available when the 27S and 19B were first released. It showed how you could use the solver to compute coordinated transforms, solve linear equations, do complex variable arithmetic, find GCD and factor prime among other things. Pretty amazing for a "non-programmable" calculator.

Yes indeed this manual is a real treasue for 27S, 17B/BII and 19B/BII owners. Reveals several undocumented (until this manual was released) features in those machines, and lots of gory examples to illustrate. Wish they had made more manuals like this one. I got one cheap on TAS last year, a steal. Also, it's on the Museum DVD; like so many other things in there, this alone is worth the price.

Check this out if you have one of these machines, and note that the Title says it's for 27S/19B only, but the 17B/BII (original Pioneer models) use the same Solver Don described. Later 17B stepchildren unfortunately have the Solve which has several issues, described by Don in earlier Forum posts (easy to find w/search).

[Katie Wasserman]

(06-10-2014 04:06 PM)Thomas Radtke Wrote:  

(06-10-2014 03:52 PM)Katie Wasserman Wrote:  Pretty amazing for a "non-programmable" calculator.

It is programmable, the solver is just not turing complete. Calculators like the 20S play in the same league, allthough not as obvious.

"not Turing Complete", I think not!

Since no calculator has an infinite memory you could say this about all of them. It's been a while since I was in grad school for this but I'm pretty sure that the solver can compute anything that a Turning machine can given enough memory. You might say that the solver has no ability to halt when looping, but it can simulate that.

[Dave Britten]

(06-10-2014 03:52 PM)Katie Wasserman Wrote:  

(06-10-2014 11:18 AM)Don Shepherd Wrote:  In my opinion, one of HP's greatest achievements was the inclusion of programming constructs such as IF, sigma (loop), and L() and G() in the solver found in several calculator models, including the 27s. This made the solver more than just an equation solver; it is a real minimum programming language.

HP should have received some type of award for this solver, but we are all beneficiaries.

If I recall correctly the HP "step-by-step" publication called HP-27S/19B Technical Applications introduced L(), G() and other solver techniques. This was available when the 27S and 19B were first released. It showed how you could use the solver to compute coordinated transforms, solve linear equations, do complex variable arithmetic, find GCD and factor prime among other things. Pretty amazing for a "non-programmable" calculator.

Nice! I hadn't spotted that one yet while browsing the DVD index. I'll have to try some of it on my 200LX.

EDIT: Interesting reading, though most of the applications are a lot more complex than an imperative program, and not terribly intuitive for the end-user. They even managed to port my favorite prime factoring algorithm, though the implementation is a bit absurd. It's extremely cool that the solver is capable of all of that, but it's clearly not the optimal choice for a complex solution. Keep a 20S handy if you want an algebraic programmable!

[Don Shepherd]

(06-10-2014 08:46 PM)Dave Britten Wrote:  They even managed to port my favorite prime factoring algorithm

Dave, here is a simple prime factoring program for the 17b/17bii solver that I wrote a few years ago. It is a lot faster than the one in the manual because it terminates the loop early by doing a divide by 0. Enter the number to factor and press N, then press FACT. If it beeps and says "Solution Not Found", it has found a factor; RCL FACT to see it. Then press FACT again to see the next (and subsequent) factors. If it doesn't beep, it has found the final factor.

It's not elegant, but it works!

Code:


IF(MOD(N:2)=0:L(FACT:2)+L(N:N/2)/0:SIGMA(I:3:SQRT(N):2:IF(MOD(N:I)=0:L(FACT:I)+L(N:N/I)/0:0))+N)-FACT

[Bruce Larrabee]

I was given my 27S at HP. No serial number, no warranty. I asked if I could have a manual and was told I wouldn't need one. After I tried the solver I agreed that I didn't. It was my main calculator on the job (programming) for years.

[Dave Britten]

(06-10-2014 10:38 PM)Don Shepherd Wrote:  

(06-10-2014 08:46 PM)Dave Britten Wrote:  They even managed to port my favorite prime factoring algorithm

Dave, here is a simple prime factoring program for the 17b/17bii solver that I wrote a few years ago. It is a lot faster than the one in the manual because it terminates the loop early by doing a divide by 0. Enter the number to factor and press N, then press FACT. If it beeps and says "Solution Not Found", it has found a factor; RCL FACT to see it. Then press FACT again to see the next (and subsequent) factors. If it doesn't beep, it has found the final factor.

It's not elegant, but it works!

Code:


IF(MOD(N:2)=0:L(FACT:2)+L(N:N/2)/0:SIGMA(I:3:SQRT(N):2:IF(MOD(N:I)=0:L(FACT:I)+L(N:N/I)/0:0))+N)-FACT

EDIT: Never mind, I found my error. Works nice!

[rprosperi]

(06-10-2014 10:38 PM)Don Shepherd Wrote:  

(06-10-2014 08:46 PM)Dave Britten Wrote:  They even managed to port my favorite prime factoring algorithm

Dave, here is a simple prime factoring program for the 17b/17bii solver that I wrote a few years ago. It is a lot faster than the one in the manual because it terminates the loop early by doing a divide by 0. Enter the number to factor and press N, then press FACT. If it beeps and says "Solution Not Found", it has found a factor; RCL FACT to see it. Then press FACT again to see the next (and subsequent) factors. If it doesn't beep, it has found the final factor.

It's not elegant, but it works!

Code:


IF(MOD(N:2)=0:L(FACT:2)+L(N:N/2)/0:SIGMA(I:3:SQRT(N):2:IF(MOD(N:I)=0:L(FACT:I)+L(N:N/I)/0:0))+N)-FACT

And the bonus with using this solver is it not only works quite well, the source code makes even hardened RPL authors do a double-take and mutter "what the $*#& is that?!" It looks like random typing, but there are just too many colons to be random...

[Katie Wasserman]

(06-10-2014 08:46 PM)Dave Britten Wrote:  Nice! I hadn't spotted that one yet while browsing the DVD index. I'll have to try some of it on my 200LX.

If you want to play with the solver on the 200LX, you can make use of the huge amount of memory for reading/writing using the 1-2-3 linking functions STOCELL and RCLCELL. Here's a gigantic equation I wrote some time ago doing that to solve for pi to a large number of decimal places (I forget what the limit is but many thousands). To use it just open 1-2-3 name a range call OUT and run this function in the solver, enter n -- the number of decimal places you want and hit CALC_PI.

Code:



{!calculate Pi to n places into OUT in 1-2-3!

calc_pi=

l(n,10000000000)*0+ !set the number of digits/cell!

sigma(x,1,length(out),1, !zero the output range! stocell(0,out,x,1)+stocell(0,out,x,2) ) +

stocell(.28*g(n),out,1,1) + !initial value for 1st term!

SIGMA(x,length(out),1+l(c,0),-1, !add to result! STOCELL(l(t,g(c)+RCLCELL(out,x,1)+RCLCELL(out,x,2))- l(c,idiv( g(t),g(n)))*g(n),out,x,2)) +

sigma(i,2,2*log(g(n))*length(out)/log(50),2, !set up 1st term loop!

SIGMA(x,l(c,0)+LENGTH(out),1, -1, !multiply by i! STOCELL(mod((g(c)+i*RCLCELL(out,x,1)),g(n))+ 0*l(c,idiv((g(c)+i*RCLCELL(out,x,1)),g(n))), out,x,1)) +

SIGMA(x,1+l(c,0),LENGTH(out),1, !divide by 50*(i+1)! STOCELL(idiv((g(c)*g(n)+RCLCELL(out,x,1)),(i+1)*50) + 0*l(c,mod((g(c)*g(n)+RCLCELL(out,x,1)), (i+1)*50)),out,x,1)) +

SIGMA(x,Length(out),1+l(c,0),-1, !add to result! STOCELL(l(t,g(c)+RCLCELL(out, x,1)+RCLCELL(out,x,2)) - l(c,idiv(g(t),g(n)))*g(n) ,out, x,2))) +

stocell(.030336*g(n),out,1,1) + !initial value for 2nd term!

SIGMA(x,length(out),1+l(c,0),-1, !add to result! STOCELL(l(t,g(c)+RCLCELL(out,x,1)+RCLCELL(out,x,2))- l(c,idiv( g(t),g(n)) ) * g( n) ,out,x,2)) +

SIGMA(i,2,2*log(g(n))*length(out)/log(6250/9),2, !setup 2nd term loop!

SIGMA(x,l(c,0)+LENGTH(out),1, -1, !multiply by i*9! STOCELL(mod((g(c)+i*9*RCLCELL(out,x,1)),g(n)) + 0*l( c, idiv((g(c)+i*9*RCLCELL(out,x,1)), g(n))),out,x,1)) +

SIGMA(x,1+l(c,0),LENGTH(out),1, !divide by 50*(i+1)! STOCELL(idiv((g(c)*g(n)+RCLCELL(out,x,1)),(i+1)*50) + 0*l(c,mod((g(c)*g(n)+RCLCELL(out,x,1)), (i+1)*50)),out,x,1)) +

SIGMA(x,1+l(c,0),LENGTH(out),1, !divide by 125! STOCELL(idiv((g(c)*g(n)+RCLCELL(out,x,1)),125) + 0*l( c,mod((g(c)*g(n)+RCLCELL(out,x,1)),125)), out, x,1)) +

SIGMA(x,LENGTH(out),1+l(c,0),-1, !add to result! STOCELL(l(t,g(c)+RCLCELL(out,x,1)+RCLCELL(out,x,2))- l(c,idiv( g(t),g(n)) ) * g( n) ,out,x,2)) ) }

Edit: Gee that code block looks terrible on this forum. Here it is on the old forum.